BULLETIN (New Series) OF THE

AMERICAN MATHEMATICAL SOCIETY

Volume 37, Number 1, Pages 5564
S 0273-0979(99)00802-2
Article electronically published on December 21, 1999
INTUITIONISM AND FORMALISM
DR. L. E. J. BROUWER
The subject for which I am asking your attention deals with the foundations of
mathematics. To understand the development of the opposing theories existing in
this eld one must rst gain a clear understanding of the concept science; for it
is as a part of science that mathematics originally took its place in human thought.
By science we mean the systematic cataloguing by means of laws of nature of
causal sequences of phenomena, i. e., sequences of phenomena which for individual
or social purposes it is convenient to consider as repeating themselves identically,
and more particularly of such causal sequences as are of importance in social rela-
tions.
That science lends such great power to man in his action upon nature is due to
the fact that the steadily improving cataloguing of ever more causal sequences of
phenomena gives greater and greater possibility of bringing about desired phenom-
ena, dicult or impossible to evoke directly, by evoking other phenomena connected
with the rst by causal sequences. And that man always and everywhere creates
order in nature is due to the fact that he not only isolates the causal sequences
of phenomena (i. e., he strives to keep them free from disturbing secondary phe-
nomena) but also supplements them with phenomena caused by his own activity,
thus making them of wider applicability. Among the latter phenomena the results
of counting and measuring take so important a place, that a large number of the
natural laws introduced by science treat only of the mutual relations between the
results of counting and measuring. It is well to notice in this connection that a
natural law in the statement of which measurable magnitudes occur can only be
understood to hold in nature with a certain degree of approximation; indeed natural
laws as a rule are not proof against sucient renement of the measuring tools.
The exceptions to this rule have from ancient times been practical arithmetic and
geometry on the one hand, and the dynamics of rigid bodies and celestial mechanics
on the other hand. Both these groups have so far resisted all improvements in the
tools of observation. But while this has usually been looked upon as something ac-
cidental and temporal for the latter group, and while one has always been prepared
to see these sciences descend to the rank of approximate theories, until compar-
atively recent times there has been absolute condence that no experiment could
ever disturb the exactness of the laws of arithmetic and geometry; this condence
is expressed in the statement that mathematics is the exact science.
On what grounds the conviction of the unassailable exactness of mathematical
laws is based has for centuries been an object of philosophical research, and two
2000 Mathematics Subject Classication. Primary 03-03, 01-02, 03Fxx.
Reprinted from Bull. Amer. Math. Soc. 20 (1913), 8196.
Translated for the Bulletin by Professor Arnold Dresden.
Inaugural address at the University of Amsterdam, read October 14, 1912.
c 1999 American Mathematical Society
55
56 DR. L. E. J. BROUWER
points of view may here be distinguished, intuitionism (largely French) and for-
malism (largely German). In many respects these two viewpoints have become
more and more denitely opposed to each other; but during recent years they have
reached agreement as to this, that the exact validity of mathematical laws as laws
of nature is out of the question. The question where mathematical exactness does
exist, is answered dierently by the two sides; the intuitionist says: in the human
intellect, the formalist says: on paper.
In Kant we nd an old form of intuitionism, now almost completely abandoned,
in which time and space are taken to be forms of conception inherent in human
reason. For Kant the axioms of arithmetic and geometry were synthetic a priori
judgments, i. e., judgments independent of experience and not capable of analyt-
ical demonstration; and this explained their apodictic exactness in the world of
experience as well as in abstracto. For Kant, therefore, the possibility of disproving
arithmetical and geometrical laws experimentally was not only excluded by a rm
belief, but it was entirely unthinkable.
Diametrically opposed to this is the view of formalism, which maintains that
human reason does not have at its disposal exact images either of straight lines
or of numbers larger than ten, for example, and that therefore these mathematical
entities do not have existence in our conception of nature any more than in nature
itself. It is true that from certain relations among mathematical entities, which
we assume as axioms, we deduce other relations according to xed laws, in the
conviction that in this way we derive truths from truths by logical reasoning, but
this non-mathematical conviction of truth or legitimacy has no exactness whatever
and is nothing but a vague sensation of delight arising from the knowledge of the
ecacy of the projection into nature of these relations and laws of reasoning. For
the formalist therefore mathematical exactness consists merely in the method of
developing the series of relations, and is independent of the signicance one might
want to give to the relations or the entities which they relate. And for the consistent
formalist these meaningless series of relations to which mathematics are reduced
have mathematical existence only when they have been represented in spoken or
written language together with the mathematical-logical laws upon which their
development depends, thus forming what is called symbolic logic.
Because the usual spoken or written languages do not in the least satisfy the
requirements of consistency demanded of this symbolic logic, formalists try to avoid
the use of ordinary language in mathematics. How far this may be carried is shown
by the modern Italian school of formalists, whose leader, Peano, published one of his
most important discoveries concerning the existence of integrals of real dierential
equations in the Mathematische Annalen in the language of symbolic logic; the
result was that it could only be read by a few of the initiated and that it did not
become generally available until one of these had translated the article into German.
The viewpoint of the formalist must lead to the conviction that if other symbolic
formulas should be substituted for the ones that now represent the fundamental
mathematical relations and the mathematical-logical laws, the absence of the sen-
sation of delight, called consciousness of legitimacy, which might be the result of
such substitution would not in the least invalidate its mathematical exactness. To
the philosopher or to the anthropologist, but not to the mathematician, belongs the
task of investigating why certain systems of symbolic logic rather than others may
be eectively projected upon nature. Not to the mathematician, but to the psychol-
ogist, belongs the task of explaining why we believe in certain systems of symbolic
INTUITIONISM AND FORMALISM 57
logic and not in others, in particular why we are averse to the so-called contradic-
tory systems in which the negative as well as the positive of certain propositions
are valid.
1
As long as the intuitionists adhered to the theory of Kant it seemed that the
development of mathematics in the nineteenth century put them in an ever weaker
position with regard to the formalists. For in the rst place this development
showed repeatedly how complete theories could be carried over from one domain
of mathematics to another: projective geometry, for example, remained unchanged
under the interchange of the r oles of point and straight line, an important part of
the arithmetic of real numbers remained valid for various complex number elds and
nearly all the theorems of elementary geometry remained true for non-archimedian
geometry, in which there exists for every straight line segment another such segment,
innitesimal with respect to the rst. These discoveries seemed to indicate indeed
that of a mathematical theory only the logical form was of importance and that
one need no more be concerned with the material than it is necessary to think of
the signicance of the digit groups with which one operates, for the correct solution
of a problem in arithmetic.
But the most serious blow for the Kantian theory was the discovery of non-
euclidean geometry, a consistent theory developed from a set of axioms diering
from that of elementary geometry only in this respect that the parallel axiom was
replaced by its negative. For this showed that the phenomena usually described
in the language of elementary geometry may be described with equal exactness,
though frequently less compactly in the language of non-euclidean geometry; hence
it is not only impossible to hold that the space of our experience has the properties
of elementary geometry but it has no signicance to ask for the geometry which
would be true for the space of our experience. It is true that elementary geometry
is better suited than any other to the description of the laws of kinematics of rigid
bodies and hence of a large number of natural phenomena, but with some patience
it would be possible to make objects for which the kinematics would be more easily
interpretable in terms of non-euclidean than in terms of euclidean geometry.
2
However weak the position of intuitionism seemed to be after this period of math-
ematical development, it has recovered by abandoning Kants apriority of space but
adhering the more resolutely to the apriority of time. This neo-intuitionism con-
siders the falling apart of moments of life into qualitatively dierent parts, to be
reunited only while remaining separated by time as the fundamental phenomenon
of the human intellect, passing by abstracting from its emotional content into the
fundamental phenomenon of mathematical thinking, the intuition of the bare two-
oneness. This intuition of two-oneness, the basal intuition of mathematics, creates
not only the numbers one and two, but also all nite ordinal numbers, inasmuch
as one of the elements of the two-oneness may be thought of as a new two-oneness,
which process may be repeated indenitely; this gives rise still further to the small-
est innite ordinal number . Finally this basal intuition of mathematics, in which
the connected and the separate, the continuous and the discrete are united, gives
rise immediately to the intuition of the linear continuum, i. e., of the between,
which is not exhaustible by the interposition of new units and which therefore can
never be thought of as a mere collection of units.
1
See Mannoury, Methodologisches und Philosophisches zur Elementarmathematik, pp. 149
154.
2
See Poincare, La Science et lHypoth`ese, p. 104.
58 DR. L. E. J. BROUWER
In this way the apriority of time does not only qualify the properties of arithmetic
as synthetic a priori judgments, but it does the same for those of geometry, and
not only for elementary two- and three-dimensional geometry, but for non-euclidean
and n-dimensional geometries as well. For since Descates we have learned to reduce
all these geometries to arithmetic by means of the calculus of coordinates.
From the present point of view of intuitionism therefore all mathematical sets of
units which are entitled to the name can be developed out of the basal intuition,
and this can only be done by combining a nite number of times the two operations:
to create a nite ordinal number and to create the innite ordinal number ;
here it is to be understood that for the latter purpose any previously constructed
set or any previously performed constructive operation may be taken as a unit.
Consequently the intuitionist recognizes only the existence of denumerable sets,
i. e., sets whose elements may be brought into one-to-one correspondence either
with the elements of a nite ordinal number or with those of the innite ordinal
number . And in the construction of these sets neither the ordinary language
nor any symbolic language can have any other r ole than that of serving as a non-
mathematical auxiliary, to assist the mathematical memory or to enable dierent
individuals to build up the same set.
For this reason the intuitionist can never feel assured of the exactness of a math-
ematical theory by such guarantees as the proof of its being non-contradictory,
3
the
possibility of dening its concepts by a nite number of words,
3
or the practical
certainty that it will never lead to a misunderstanding in human relations.
4
As has been stated above, the formalist wishes to leave to the psychologist the
task of selecting the truly-mathematical language from among the many symbolic
languages that may be consistently developed. Inasmuch as psychology has not yet
begun on this task, formalism is compelled to mark o, at least temporarily, the
domain that it wishes to consider as true mathematics and to lay down for that
purpose a denite system of axioms and laws of reasoning, if it does not wish
to see its work doomed to sterility. The various ways in which this attempt has
actually been made all follow the same leading idea, viz., the presupposition of the
existence of a world of mathematical objects, a world independent of the thinking
individual, obeying the laws of classical logic and whose objects may possess with
respect to each other the relation of a set to its elements. With reference to
this relation various axioms are postulated, suggested by the practice with natural
nite sets; the principal of these are: a set is determined by its elements; for
any two mathematical objects it is decided whether or not one of them is contained
in the other one as an element ; to every set belongs another set containing as
its elements nothing but the subsets of the given set ; the axiom of selection: a
set which is split into subsets contains at least one subset which contains one and
not more than one element of each of the rst subsets; the axiom of inclusion: if
for any mathematical object it is decided whether a certain property is valid for it
or not, then there exists a set containing nothing but those objects for which the
property does hold; the axiom of composition: the elements of all sets that belong
to a set of sets form a new set.
On the basis of such a set of axioms the formalist develops now in the rst
place the theory of nite sets. A set is called nite if its elements can not be
3
See however Poincare in Scientia, No. XXIV, p. 6.
4
See however Borel in Revue du Mois, No. 80, p. 221.
INTUITIONISM AND FORMALISM 59
brought into one-to-one correspondence with the elements of one of its subsets;
by means of relatively complicated reasoning the principle of complete induction
is proved to be a fundamental property of these sets;
5
this principle states that a
property will be true for all nite sets if, rst, it is true for all sets containing a
single element, and, second, its validity for an arbitrary nite set follows from its
validity for this same set reduced by a single one of its elements. That the formalist
must give an explicit proof of this principle, which is self-evident for the nite
numbers of the intuitionist on account of their construction, shows at the same
time that the former will never be able to justify his choice of axioms by replacing
the unsatisfactory appeal to inexact practice or to intuition equally inexact for him
by a proof of the non-contradictoriness of his theory. For in order to prove that a
contradiction can never arise among the innitude of conclusions that can be drawn
from the axioms he is using, he would rst have to show that if no contradiction
had as yet arisen with the nth conclusion then none could arise with the (n +1)th
conclusion, and secondly he would have to apply the principle of complete induction
intuitively. But it is this last step which the formalist may not take, even though
he should have proved the principle of complete induction; for this would require
mathematical certainty that the set of properties obtained after the nth conclusion
had been reached, would satisfy for an arbitrary n his denition for nite sets,
6
and in order to secure this certainly he would have to have recourse not only to the
unpermissible application of a symbolic criterion to a concrete example but also to
another intuitive application of the principle of complete induction; this would lead
him to a vicious circle reasoning.
In the domain of nite sets in which the formalistic axioms have an interpreta-
tion perfectly clear to the intuitionists, unreservedly agreed to by them, the two
tendencies dier solely in their method, not in their results; this becomes quite
dierent however in the domain of innite or transnite sets, where, mainly by the
application of the axiom of inclusion, quoted above, the formalist introduces vari-
ous concepts, entirely meaningless to the intuitionist, such as for instance the set
whose elements are the points of space, the set whose elements are the continuous
functions of a variable, the set whose elements are the discontinuous functions of
a variable, and so forth. In the course of these formalistic developments it turns
out that the consistent application of the axiom of inclusion leads inevitably to
contradictions. A clear illustration of this fact is furnished by the so-called paradox
of Burali-Forti.
7
To exhibit it we have to lay down a few denitions.
A set is called ordered if there exists between any two of its elements a relation
of higher than or lower than, with this understanding that if the element a is
higher than the element b, then the element b is lower than the element a, and if
the element b is higher than a and c is higher than b, then c is higher than a.
A well-ordered set (in the formalistic sense) is an ordered set, such that every
subset contains an element lower than all others.
Two well-ordered sets that may be brought into one-to-one correspondence under
invariance of the relations of higher than and lower than are said to have the
same ordinal number.
5
Compare, e. g., Zermelo, Sur les ensembles nis et le principe de linduction compl`ete, Acta
Mathematica, 32, pp. 185193.
6
Compare Poincare, Revue de Metaphysique et de Morale, 1905, p. 834.
7
Compare Rendiconti del Circolo Matematico di Palermo, 1897.
60 DR. L. E. J. BROUWER
If two ordinal numbers A and B are not equal, then one of them is greater than
the other one, let us say A is greater than B; this means that B may be brought
into one-to-one correspondence with an initial segment of A under invariance of
the relations of higher than and lower than. We have introduced above, from
the intuitionist viewpoint, the smallest innite ordinal number , i. e., the ordinal
number of the set of all nite ordinal numbers arranged in order of magnitude.
8
Well-ordered sets having the ordinal number are called elementary series.
It is proved without diculty by the formalist that an arbitrary subset of a well-
ordered set is also a well-ordered set, whose ordinal number is less than or equal to
that of the original set; also, that if to a well-ordered set that does not contain all
mathematical objects a new element be added that is dened to be higher than all
elements of the original set, a new well-ordered set arises whose ordinal number is
greater than that of the rst set.
We construct now on the basis of the axiom of inclusion the set s which contains
as elements all the ordinal numbers arranged in order of magnitude; then we can
prove without diculty, on the one hand that s is a well-ordered set whose ordinal
number can not be exceeded by any other ordinal number in magnitude, and on
the other hand that it is possible, since not all mathematical objects are ordinal
numbers, to create an ordinal number greater than that of s by adding a new
element to s, a contradiction.
9
Although the formalists must admit contradictory results as mathematical if
they want to be consistent, there is something disagreeable for them in a paradox
like that of Burali-Forti because at the same time the progress of their arguments is
guided by the principium contradictionis, i. e., by the rejection of the simultaneous
validity of two contradictory properties. For this reason the axiom of inclusion has
been modied to read as follows: If for all elements of a set it is decided whether a
certain property is valid for them or not, then the set contains a subset containing
nothing but those elements for which the property does hold.
10
In this form the axiom permits only the introduction of such sets as are subsets of
sets previously introduced; if one wishes to operate with other sets, their existence
must be explicitly postulated. Since however in order to accomplish anything at all
the existence of a certain collection of sets will have to be postulated at the outset,
the only valid argument that can be brought against the introduction of a new set is
that it leads to contradictions; indeed the only modications that the discovery of
paradoxes has brought about in the practice of formalism has been the abolition of
those sets that had given rise to these paradoxes. One continues to operate without
hesitation with other sets introduced on the basis of the old axiom of inclusion; the
result of this is that extended elds of research, which are without signicance for
the intuitionist are still of considerable interest to the formalist. An example of this
8
The more general ordinal numbers of the intuitionist are the numbers constructed by means
of Cantors two principles of generation (compare Math. Annalen, vol. 49, p. 226).
9
It is without justice that the paradox of Burali-Forti is sometimes classed with the of Richard,
which in a somewhat simplied form reads as follows: Does there exist a least integer, that can
not be dened by a sentence of at most twenty words? On the one hand yes, for the number of
sentences of at most twenty words is of course nite; on the other hand no, for if it should exist,
it would be dened by the sentence of fteen words formed by the words italicized above.
The origin of this paradox does not lie in the axiom of inclusion but in the variable meaning of
the word dened in the italicized sentence, which makes it possible to dene by means of this
sentence an innite number of integers in succession.
10
Compare Zermelo, Math. Annalen, vol. 65, p. 263.
INTUITIONISM AND FORMALISM 61
is found in the theory of potencies, of which I shall sketch the principal features
here, because it illustrates so clearly the impassable chasm which separates the two
sides.
Two sets are said to possess the same potency, or power, if their elements can be
brought into one-to-one correspondence. The power of set A is said to be greater
than that of B, and the power of B less than that of A, if it is possible to establish a
one-to-one correspondence between B and a part of A, but impossible to establish
such a correspondence between A and a part of B. The power of a set which has the
same power as one of its subsets, is called innite, other powers are called nite.
Sets that have the same power as the ordinal number are called denumerably
innite and the power of such sets is called aleph-null: it proves to be the smallest
innite power. According to the statements previously made, this power aleph-null
is the only innite power of which the intuitionists recognize the existence.
Let us now consider the concept: denumerably innite ordinal number. From
the fact that this concept has a clear and well-dened meaning for both formalist
and intuitionist, the former infers the right to create the set of all denumerably
innite ordinal numbers, the power of which he calls aleph-one, a right not rec-
ognized by the intuitionist. Because it is possible to argue to the satisfaction of
both formalist and intuitionist, rst, that denumerably innite sets of denumerably
innite ordinal numbers can be built up in various ways, and second, that for every
such set it is possible to assign a denumerably innite ordinal number, not be-
longing to this set, the formalist concludes: aleph-one is greater than aleph-null,
a proposition, that has no meaning for the intuitionist. Because it is possible to
argue to the satisfaction of both formalist and intuitionist that it is impossible to
construct
11
a set of denumerably innite ordinary numbers, which could be proved
to have a power less than that of aleph-one, but greater than that of aleph-null, the
formalist concludes: aleph-one is the second smallest innite ordinal number, a
proposition that has no meaning for the intuitionist.
Let us consider the concept: real number between 0 and 1. For the formalist
this concept is equivalent to elementary series of digits after the decimal point,
12
for the intuitionist it means law for the construction of an elementary series of
digits after the decimal point, built up by means of a nite number of operations.
And when the formalist creates the set of all real numbers between 0 and 1,
these words are without meaning for the intuitionist, even whether one thinks of
the real numbers of the formalist, determined by elementary series of freely selected
digits, or of the real numbers of the intuitionist, determined by elementary series
of freely selected digits, or of the real numbers of the intuitionist, determined by
nite laws of construction. Because it is possible to prove to the satisfaction of
both formalist and intuitionist, rst, that denumerably innite sets of real numbers
between 0 and 1 can be constructed in various ways, and second that for every such
set it is possible to assign a real number between 0 and 1, not belonging to the set,
the formalist concludes: the power of the continuum, i. e., the power of the set
of real numbers between 0 and 1, is greater than aleph-null, a proposition which
11
If construct were here replaced by dene (in the formalistic sense), the proof would not
be satisfactory to the intuitionist. For, in Cantors argument in Math. Annalen, vol. 49, it is not
allowed to replace the words konnen wir bestimmen (p. 214, line 17 from top) by the words
muss es geben.
12
Here as everywhere else in this paper, the assumption is tacitly made that there are an
innite number of digits dierent from 9.
62 DR. L. E. J. BROUWER
is without meaning for the intuitionist; the formalist further raises the question,
whether there exist sets of real numbers between 0 and 1, whose power is less than
that of the continuum, but greater than aleph-null, in other words, whether the
power of the continuum is the second smallest innite power, and this question,
which is still waiting for an answer, he considers to be one of the most dicult and
most fundamental of mathematical problems.
For the intuitionist, however, the question as stated is without meaning; and as
soon as it has been so interpreted as to get a meaning, it can easily be answered.
If we restate the question in this form: Is it impossible to construct
13
innite sets
of real numbers between 0 and 1, whose power is less than that of the continuum,
but greater than aleph-null?, then the answer must be in the armative; for the
intuitionist can only construct denumerable sets of mathematical objects and if, on
the basis of the intuition of the linear continuum, he admits elementary series of free
selections as elements of construction, then each non-denumerable set constructed
by means of it contains a subset of the power of the continuum.
If we restate the question in the form: Is it possible to establish a one-to-
one correspondence between the elements of a set of denumerably innite ordinal
numbers on the one hand, and a set of real numbers between 0 and 1 on the other
hand, both sets being indenitely extended by the construction of new elements, of
such a character that the correspondence shall not be disturbed by any continuation
of the construction of both sets?, then the answer must also be in the armative,
for the extension of both sets can be divided into phases in such a way as to add a
denumerably innite number of elements during each phase.
14
If however we put the question in the following form: Is it possible to construct
a law which will assign a denumerably innite ordinal number to every elementary
series of digits and which will give certainty a priori that two dierent elementary
series will never have the same denumerably innite ordinal number corresponding
to them?, then the answer must be in the negative; for this law of correspon-
dence must prescribe in some way a construction of certain denumerably innite
ordinal numbers at each of the successive places of the elementary series; hence
there is for each place c

a well-dened largest denumerably innite number

,
the construction of which is suggested by that particular place; there is then also a
well-dened denumerably innite ordinal number

, greater than all

s and that
can not therefore be exceeded by any of the ordinal numbers involved by the law
of correspondence; hence the power of that set of ordinal numbers can not exceed
aleph-null.
As a means for obtaining ever greater powers, the formalists dene with every
power a set of all the dierent ways in which a number of selections of power
may be made, and they prove that the power of this set is greater that .
13
If construct were here replaced by dene (in the formalistic sense), and if we suppose
that the problem concerning the pairs of digits in the decimal fraction development of , discussed
on p. 95, can not be solved, then the question of the text must be answered negatively. For, let
us denote by Z the set of those innite binary fractions, whose nth digit is 1, if the nth pair of
digits in the decimal fraction development of consists of unequal digits; let us further denote by
X the set of all nite binary fractions. Then the power of Z + X is greater than aleph-null, but
less than that of the continuum.
14
Calling denumerably unnished all sets of which the elements can be individually realized,
and in which for every denumerably innite subset there exists an element not belonging to this
subset, we can say in general, in accordinance with the denitions of the text: All denumerably
unnished sets have the same power.
INTUITIONISM AND FORMALISM 63
In particular, when it has been proved to the satisfaction of both formalist and
intuitionist that it is possible in various ways to construct laws according to which
functions of a real variable dierent from each other are made to correspond to all
elementary series of digits, but that it is impossible to construct a law according
to which an elementary series of digits is made to correspond to every function of
a real variable and in which there is certainty a priori that two dierent functions
will never have the same elementary series corresponding to them, the formalist
concludes: the power c

of the set of all functions of a real variable is greater than

the power c of the continuum, a proposition without meaning to the intuitionist;
and in the same way in which he was led from c to c

, he comes from c

to a still
greater power c

.
A second method used by the formalists for obtaining ever greater powers is to
dene for every power , which can serve as a power of ordinary numbers, the set
of all ordinal numbers of power , and then to prove that the power of this set
is greater than . In particular they denote by aleph-two the power of the set of
all ordinal numbers of power aleph-one and they prove that aleph-two is greater
than aleph-one and that it follows in magnitude immediately after aleph-one. If it
should be possible to interpret this result in a way in which it would have meaning
for the intuitionist, such interpretation would not be as simple in this case as it was
in the preceding cases.
What has been treated so far must be considered to be the negative part of
the theory of potencies; for the formalist there also exists a positive part however,
founded on the theorem of Bernstein: If the set A has the same power as a subset
of B and B has the same power as a subset of A, then A and B have the same
power or, in an equivalent form: If the set A = A
1
+B
1
+C
1
, has the same power
as the set A
1
, then it also has the same power as the set A
1
+B
1
.
This theorem is self-evident for denumerable sets. If it is to have any meaning
at all for sets of higher power for the intuitionist, it will have to be interpretable
as follows: If it is possible, rst to construct a law determining a one-to-one
correspondence between the mathematical entities of type A and those of type A
1
and second to construct a law determining a one-to-one correspondence between the
mathematical entities of type A and those of types A
1
, B
1
and C
1
, then it is possible
to determine from these two laws by means of a nite number of operations a third
law, determining a one-to-one correspondence between the mathematical entities
of type A and those of types A
1
and B
1
.
In order to investigate the validity of this interpretation, we quote the proof:
From the division of A in to A
1
+ B
1
+ C
1
, we secure by means of the corre-
spondence
1
between A and A
1
a division of A
1
into A
2
+ B
2
+ C
2
, as well as a
one-to-one correspondence
2
between A
1
and A
2
. From the division of A
1
into
A
2
+B
2
+C
2
, we secure by means of the correspondence between A
1
and A
2
a divi-
sion of A
2
into A
3
+B
3
+C
3
, as well as a one-to-one correspondence
3
between A
2
and A
3
. Indenite repetition of this procedure will divide the set A into an elemen-
tary series of subsets C
1
, C
2
, C
3
, . . . , an elementary series of subsets B
1
, B
2
, B
3
. . . ,
and a remainder set D. The correspondence
C
between A and A
1
+ B
1
which is
desired is secured by assigning to every element of C

the corresponding element

of C
+1
and by assigning every other element of A to itself.
In order to test this proof on a denite example, let us take for A the set of all
real numbers between 0 and 1, represented by innite decimal fractions, for A
1
the
set of those decimal fractions in which the (2n1)th digit is equal to the 2nth digit;
64 DR. L. E. J. BROUWER
further a decimal fraction that does not belong to A
1
will be counted to belong to
B
1
or to C
1
, according as the above-mentioned equality of digits occurs an innite
or a nite number of times. By replacing successively each digit of an arbitrary
element of A by a pair of digits equal to it, we secure at once a law determining
a one-to-one correspondence
1
between A and A
1
. For the element of A
1
that
corresponds to an arbitrary well-dened element of A, such as, e. g., 3, we can
determine successively as many digits as we please; it must therefore be considered
as being well-dened.
In order to determine the element corresponding to 3 according to the cor-
respondence
C
, it is now necessary to decide rst whether it happens an innite
or a nite number of times in the decimal fraction development of 3 that a
digit in an odd-numbered place is equal to the digit in the following even-numbered
place; for this purpose we should either have to invent a process for constructing
an elementary series of such pairs of equal digits, or to deduce a contradiction from
the assumption of the existence of such an elementary series. There is, however, no
ground for believing that either of these problems can be solved.
Hence it has become evident that also the theorem of Bernstein, and with it the
positive part of the theory of potencies, does not allow an intuitionistic interpreta-
tion.
So far my exposition of the fundamental issue, which divides the mathematical
world. There are eminent scholars on both sides and the chance of reaching an
agreement within a nite period is practically excluded. To speak with Poincare:
Les hommes ne sentendent pas, parce quils ne parlent pas la meme langue et
quil y a des langues qui ne sapprennent pas.